linux-kernel - Re: [PATCH 2/3] i2c-s3c2410: Rework device type handling

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20120417173136.GB22406@pengutronix.de>
Date:	Tue, 17 Apr 2012 19:31:36 +0200
From:	Wolfram Sang <w.sang@...gutronix.de>
To:	Karol Lewandowski <k.lewandowsk@...sung.com>
Cc:	ben-linux@...ff.org, m.szyprowski@...sung.com,
	thomas.abraham@...aro.org, linux-kernel@...r.kernel.org,
	linux-i2c@...r.kernel.org, devicetree-discuss@...ts.ozlabs.org,
	linux-samsung-soc@...r.kernel.org, t.stanislaws@...sung.com,
	kyungmin.park@...sung.com, broonie@...nsource.wolfsonmicro.com,
	Grant Likely <grant.likely@...retlab.ca>,
	Rob Herring <robherring2@...il.com>
Subject: Re: [PATCH 2/3] i2c-s3c2410: Rework device type handling

Hi,

On Wed, Mar 21, 2012 at 08:11:52PM +0100, Karol Lewandowski wrote:
> Reorganize driver a bit to better handle device tree-based systems:
> 
>  - move machine type to driver's private structure instead of
>    quering platform device variants in runtime
> 
>  - replace s3c24xx_i2c_type enum with unsigned int that holds
>    bitmask with revision-specific quirks
> 
> Signed-off-by: Karol Lewandowski <k.lewandowsk@...sung.com>
> Signed-off-by: Kyungmin Park <kyungmin.park@...sung.com>

Okay, so this driver needs to the 'data' field from either
platform_device_id or of_device_id and implements a function for that,
namely s3c24xx_get_device_quirks(). Grant, Rob: I'd think there might be
more drivers in need of that, so maybe it makes sense to have a generic
of-helper function?

> ---
>  drivers/i2c/busses/i2c-s3c2410.c |   47 ++++++++++++++++++-------------------
>  1 files changed, 23 insertions(+), 24 deletions(-)
> 
> diff --git a/drivers/i2c/busses/i2c-s3c2410.c b/drivers/i2c/busses/i2c-s3c2410.c
> index 85e3664..f7b6a14 100644
> --- a/drivers/i2c/busses/i2c-s3c2410.c
> +++ b/drivers/i2c/busses/i2c-s3c2410.c
> @@ -44,8 +44,14 @@
>  #include <plat/regs-iic.h>
>  #include <plat/iic.h>
>  
> -/* i2c controller state */
> +#ifdef CONFIG_OF
> +static const struct of_device_id s3c24xx_i2c_match[];
> +#endif

Uh, forward declaration with #ifdef. I'd think we should get this simply
to the front.

>  dd
> +/* Treat S3C2410 as baseline hardware, anything else is supported via quirks */
> +#define QUIRK_S3C2440		(1 << 0)

Minor: Is it really a quirk being s3c2440? Maybe FLAG_*, dunno.

> +
> +/* i2c controller state */
>  enum s3c24xx_i2c_state {
>  	STATE_IDLE,
>  	STATE_START,
> @@ -54,14 +60,10 @@ enum s3c24xx_i2c_state {
>  	STATE_STOP
>  };
>  
> -enum s3c24xx_i2c_type {
> -	TYPE_S3C2410,
> -	TYPE_S3C2440,
> -};
> -
>  struct s3c24xx_i2c {
>  	spinlock_t		lock;
>  	wait_queue_head_t	wait;
> +	unsigned int            quirks;
>  	unsigned int		suspended:1;
>  
>  	struct i2c_msg		*msg;
> @@ -88,26 +90,22 @@ struct s3c24xx_i2c {
>  #endif
>  };
>  
> -/* default platform data removed, dev should always carry data. */
> -
> -/* s3c24xx_i2c_is2440()
> +/* s3c24xx_get_device_quirks
>   *
> - * return true is this is an s3c2440
> + * Get controller type either from device tree or platform device variant.
>  */
>  
> -static inline int s3c24xx_i2c_is2440(struct s3c24xx_i2c *i2c)
> +static inline unsigned int s3c24xx_get_device_quirks(struct platform_device *pdev)
>  {
> -	struct platform_device *pdev = to_platform_device(i2c->dev);
> -	enum s3c24xx_i2c_type type;
> -
>  #ifdef CONFIG_OF
> -	if (i2c->dev->of_node)
> -		return of_device_is_compatible(i2c->dev->of_node,
> -				"samsung,s3c2440-i2c");
> +        if (pdev->dev.of_node) {
> +		const struct of_device_id *match;
> +		match = of_match_node(&s3c24xx_i2c_match[0], pdev->dev.of_node);

Minor: I think it is more readable to drop the [0]

> +		return (unsigned int)match->data;
> +        }
>  #endif
>  
> -	type = platform_get_device_id(pdev)->driver_data;
> -	return type == TYPE_S3C2440;
> +	return platform_get_device_id(pdev)->driver_data;
>  }
>  
>  /* s3c24xx_i2c_master_complete
> @@ -676,7 +674,7 @@ static int s3c24xx_i2c_clockrate(struct s3c24xx_i2c *i2c, unsigned int *got)
>  
>  	writel(iiccon, i2c->regs + S3C2410_IICCON);
>  
> -	if (s3c24xx_i2c_is2440(i2c)) {
> +	if (i2c->quirks & QUIRK_S3C2440) {
>  		unsigned long sda_delay;
>  
>  		if (pdata->sda_delay) {
> @@ -906,6 +904,7 @@ static int s3c24xx_i2c_probe(struct platform_device *pdev)
>  		goto err_noclk;
>  	}
>  
> +	i2c->quirks = s3c24xx_get_device_quirks(pdev);
>  	if (pdata)
>  		memcpy(i2c->pdata, pdata, sizeof(*pdata));
>  	else
> @@ -1113,18 +1112,18 @@ static const struct dev_pm_ops s3c24xx_i2c_dev_pm_ops = {
>  static struct platform_device_id s3c24xx_driver_ids[] = {
>  	{
>  		.name		= "s3c2410-i2c",
> -		.driver_data	= TYPE_S3C2410,
> +		.driver_data	= 0,
>  	}, {
>  		.name		= "s3c2440-i2c",
> -		.driver_data	= TYPE_S3C2440,
> +		.driver_data	= QUIRK_S3C2440,
>  	}, { },
>  };
>  MODULE_DEVICE_TABLE(platform, s3c24xx_driver_ids);
>  
>  #ifdef CONFIG_OF
>  static const struct of_device_id s3c24xx_i2c_match[] = {
> -	{ .compatible = "samsung,s3c2410-i2c" },
> -	{ .compatible = "samsung,s3c2440-i2c" },
> +	{ .compatible = "samsung,s3c2410-i2c", .data = (void *)0 },
> +	{ .compatible = "samsung,s3c2440-i2c", .data = (void *)QUIRK_S3C2440 },

Hmm, the need to sepcify the quirks twice may lead to only one instance
being updated in future patches, but I don't see a way around that,
currently :(

>  	{},
>  };
>  MODULE_DEVICE_TABLE(of, s3c24xx_i2c_match);
> -- 
> 1.7.9
> 

Thanks,

   Wolfram

-- 
Pengutronix e.K.                           | Wolfram Sang                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

Download attachment "signature.asc" of type "application/pgp-signature" (199 bytes)